Edge Direction-invariant Graph Neural Networks for Molecular Dipole Moments Prediction
This addresses the challenge of accurately predicting molecular dipole moments for computational chemistry applications, representing a domain-specific incremental improvement over existing equivariant embeddings.
The researchers tackled the problem of predicting molecular dipole moments by developing a graph neural network that directly encodes physical implications of local dipole contributions rather than treating molecules as topological graphs. Their model achieved prediction accuracy comparable to ab-initio calculations and worked effectively even for molecules with extended geometries.
The dipole moment is a physical quantity indicating the polarity of a molecule and is determined by reflecting the electrical properties of constituent atoms and the geometric properties of the molecule. Most embeddings used to represent graph representations in traditional graph neural network methodologies treat molecules as topological graphs, creating a significant barrier to the goal of recognizing geometric information. Unlike existing embeddings dealing with equivariance, which have been proposed to handle the 3D structure of molecules properly, our proposed embeddings directly express the physical implications of the local contribution of dipole moments. We show that the developed model works reasonably even for molecules with extended geometries and captures more interatomic interaction information, significantly improving the prediction results with accuracy comparable to ab-initio calculations.